It is my pleasure to be part of GERAD, a large and diverse research center, where I can learn from and have an opportunity to collaborate with great researchers in the areas of optimization and data science. My early research, when I was a PhD student, focused on an integrated planning system in supply chains to efficiently determine the decisions in production, inventory and distribution simultaneously. The framework, later extended, to include the case of uncertain demand. My postdoctoral research served to further examined the decomposition techniques for stochastic and robust vehicle routing problems. At the same time, I have expanded my interests to multi-agent systems in artificial intelligence as well as sequential decision process under uncertainty. Before returning back to academia, I worked at JDA Innovation Labs in Montreal as a data scientist to develop data-driven tools in several applications including retail analytics and supply chain planning using various machine learning and optimization techniques.

My recent interest in research lies in the applications of decision and data analytics in supply chain management. In terms of methodologies, I am particularly interested in the algorithmic developments of large-scale stochastic and robust optimization as well as sequential decision algorithms with possible enhancements from data analytics.

One of my focuses is to enable new capabilities at the beginning of the supply chain planning process, namely demand forecast and inventory planning. I wish to explore general forms of demand uncertainty predictions that can be used to better describe non-regular items with a number of zero demand observations and lumpy patterns, and investigate how they can be embedded directly into the inventory optimization frameworks as well as subsequent planning processes.

Another relevant area of focus is the optimization of a multi-stage manufacturing and distribution plan under uncertainty in the context of material requirements planning (MRP) and distribution resource planning (DRP). In an uncertain environment, the plan determined by using deterministic optimization (where demand and lead time are assumed to be known) could become infeasible or result in poor performance. In addition to this, we would like to explore learning algorithms that can be used to anticipate disruptions and recommend a set of actions that should be put in place to neutralize impact of the disruptions. These applications would allow us to bridge the gap between planning and execution in supply chain operations.

Member of GERAD since February 2017